Data analysis for the pilot participants in PLEARN

Load Libraries and Resources¶

library('readr')
library('blabr')
library('plyr')
library('dplyr')
library('ggplot2')
library('rjson')
library('parallel')
library("readxl")
source('PLEARN_analysis_helper.R')

Attaching package: ‘dplyr’

The following objects are masked from ‘package:plyr’:

    arrange, count, desc, failwith, id, mutate, rename, summarise,
    summarize

The following objects are masked from ‘package:stats’:

    filter, lag

The following objects are masked from ‘package:base’:

    intersect, setdiff, setequal, union

item_properties = read.csv('../PLEARN_experiments/csv/plearn_receptiveTask_datasource.csv')
item_properties[1,]

# audio timings for the original version ("bg_")
bg_audio_timings = getAudioTimingsFromGlob("../PLEARN_experiments/stimuli/raw_audio2/*.cut")
bg_audio_timings_with_metadata = merge(item_properties, bg_audio_timings, by.x='AudioTarget', 
    by.y='audiotarget')
bg_audio_timings_with_metadata$plural_duration = (bg_audio_timings_with_metadata$target_noun_end_time -
    bg_audio_timings_with_metadata$begin_disambig_region) * 1000 
print(nrow(bg_audio_timings_with_metadata))

[1] 63

# audio timings for the red/blue version ("rb_")
rb_audio_timings = getAudioTimingsFromGlob("../PLEARN_experiments/stimuli/raw_audio_simplified/*.cut")
rb_audio_timings_with_metadata = merge(item_properties, rb_audio_timings, by.x='AudioTarget', 
    by.y='audiotarget')
# we don't have duration information until Chompsky is up again 
rb_audio_timings_with_metadata$plural_duration = (rb_audio_timings_with_metadata$target_noun_end_time -
    rb_audio_timings_with_metadata$begin_disambig_region) * 1000 
print(nrow(rb_audio_timings_with_metadata))

[1] 63

subject_info = read.csv('subject_info.csv', stringsAsFactors=F) #No PII
subject_info$participant_name = sapply(strsplit(subject_info$filename,'_' ), function(x){x[1]})
subject_info$age_in_months = subject_info$age_in_days / 30.5
subject_info

Audio Stimuli Properties¶

Plural Segment Duration (Background version)¶

options(repr.plot.width=4, repr.plot.height=3)
ggplot(subset(bg_audio_timings_with_metadata, plural_duration != 0)) + geom_histogram(aes(x=plural_duration)
) + facet_wrap(~voicing, ncol=1) + theme_bw() + geom_vline(xintercept=0) +xlab('duration (ms)')

`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Prepositional Phrase Duration (Background version)¶

mean_pp_duration= mean(bg_audio_timings$pp_duration, na.rm=T) * 1000
options(repr.plot.width=4, repr.plot.height=3)
ggplot(bg_audio_timings) + geom_histogram(aes(x=pp_duration * 1000)) + theme_bw() + xlab('Prepositional Phrase Duration (ms)')

`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Plural Segment Duration (Simplified version)¶

options(repr.plot.width=4, repr.plot.height=3)
ggplot(subset(rb_audio_timings_with_metadata, plural_duration > 0)) + geom_histogram(aes(x=plural_duration)
) + facet_wrap(~voicing, ncol=1) + theme_bw() + geom_vline(xintercept=0) +xlab('duration (ms)')

`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

#[ ]!!! check on the three outliers

Preprocess Eyetracking Data¶

# plz parallelize me
source('PLEARN_analysis_helper.R')
ed = list()
for (i in 1:nrow(subject_info)){
    participant = subject_info[i,]
    if (participant$filename == '' | is.null(participant$filename) | participant$include == 0){
        print(paste0('No eyetracking data for ', participant$id,' (or excluded), skipping...'))
    } else {
        if (participant$expt_version == 'scene'){
            audio_timings = bg_audio_timings
        } else if (participant$expt_version == 'redblue') {
            audio_timings = rb_audio_timings
        } else {
            stop('expt_version not recognized')
        }        
        ed[[participant$id]] = analyzeEyetrackingParticipant(
            'results/', 
            participant$filename, 
            audio_timings, 
            participant$type)
    }
}

[1] "processing results/smflip3_fixreport.txt..."

Parsed with column specification:
cols(
  .default = col_character(),
  TRIAL_INDEX = col_double(),
  CURRENT_FIX_ADJUSTED = col_logical(),
  CURRENT_FIX_DURATION = col_double(),
  CURRENT_FIX_END = col_double(),
  CURRENT_FIX_INDEX = col_double(),
  CURRENT_FIX_IS_RT_END = col_logical(),
  CURRENT_FIX_MSG_COUNT = col_double(),
  CURRENT_FIX_PUPIL = col_double(),
  CURRENT_FIX_START = col_double(),
  CURRENT_FIX_X = col_double(),
  CURRENT_FIX_X_RESOLUTION = col_double(),
  CURRENT_FIX_Y = col_double(),
  CURRENT_FIX_Y_RESOLUTION = col_double(),
  IP_END_EVENT_MATCHED = col_logical(),
  IP_END_TIME = col_double(),
  IP_START_EVENT_MATCHED = col_logical(),
  IP_START_TIME = col_double(),
  TRIAL_FIXATION_TOTAL = col_double(),
  TRIAL_START_TIME = col_double(),
  RT = col_double()
  # ... with 8 more columns
)
See spec(...) for full column specifications.
Warning message:
“Length of logical index must be 1 or 13104, not 13103”

[1] "processing results/pl00_fixations.txt..."

Parsed with column specification:
cols(
  .default = col_character(),
  TRIAL_INDEX = col_double(),
  CURRENT_FIX_ADJUSTED = col_logical(),
  CURRENT_FIX_DURATION = col_double(),
  CURRENT_FIX_END = col_double(),
  CURRENT_FIX_INDEX = col_double(),
  CURRENT_FIX_IS_RT_END = col_logical(),
  CURRENT_FIX_MSG_COUNT = col_double(),
  CURRENT_FIX_PUPIL = col_double(),
  CURRENT_FIX_START = col_double(),
  CURRENT_FIX_X = col_double(),
  CURRENT_FIX_X_RESOLUTION = col_double(),
  CURRENT_FIX_Y = col_double(),
  CURRENT_FIX_Y_RESOLUTION = col_double(),
  IP_END_EVENT_MATCHED = col_logical(),
  IP_END_TIME = col_double(),
  IP_START_EVENT_MATCHED = col_logical(),
  IP_START_TIME = col_double(),
  TRIAL_FIXATION_TOTAL = col_double(),
  TRIAL_START_TIME = col_double(),
  RT = col_double()
  # ... with 8 more columns
)
See spec(...) for full column specifications.
Warning message:
“Length of logical index must be 1 or 15018, not 15017”

[1] "processing results/pl01_fixations.txt..."

Parsed with column specification:
cols(
  .default = col_character(),
  TRIAL_INDEX = col_double(),
  CURRENT_FIX_ADJUSTED = col_logical(),
  CURRENT_FIX_DURATION = col_double(),
  CURRENT_FIX_END = col_double(),
  CURRENT_FIX_INDEX = col_double(),
  CURRENT_FIX_IS_RT_END = col_logical(),
  CURRENT_FIX_MSG_COUNT = col_double(),
  CURRENT_FIX_PUPIL = col_double(),
  CURRENT_FIX_START = col_double(),
  CURRENT_FIX_X = col_double(),
  CURRENT_FIX_X_RESOLUTION = col_double(),
  CURRENT_FIX_Y = col_double(),
  CURRENT_FIX_Y_RESOLUTION = col_double(),
  IP_END_EVENT_MATCHED = col_logical(),
  IP_END_TIME = col_double(),
  IP_START_EVENT_MATCHED = col_logical(),
  IP_START_TIME = col_double(),
  TRIAL_FIXATION_TOTAL = col_double(),
  TRIAL_START_TIME = col_double(),
  RT = col_double()
  # ... with 8 more columns
)
See spec(...) for full column specifications.
Warning message:
“Length of logical index must be 1 or 14335, not 14334”

[1] "processing results/pl03_fixations.txt..."

Parsed with column specification:
cols(
  .default = col_character(),
  TRIAL_INDEX = col_double(),
  CURRENT_FIX_ADJUSTED = col_logical(),
  CURRENT_FIX_DURATION = col_double(),
  CURRENT_FIX_END = col_double(),
  CURRENT_FIX_INDEX = col_double(),
  CURRENT_FIX_IS_RT_END = col_logical(),
  CURRENT_FIX_MSG_COUNT = col_double(),
  CURRENT_FIX_PUPIL = col_double(),
  CURRENT_FIX_START = col_double(),
  CURRENT_FIX_X = col_double(),
  CURRENT_FIX_X_RESOLUTION = col_double(),
  CURRENT_FIX_Y = col_double(),
  CURRENT_FIX_Y_RESOLUTION = col_double(),
  IP_END_EVENT_MATCHED = col_logical(),
  IP_END_TIME = col_double(),
  IP_START_EVENT_MATCHED = col_logical(),
  IP_START_TIME = col_double(),
  TRIAL_FIXATION_TOTAL = col_double(),
  TRIAL_START_TIME = col_double()
)
See spec(...) for full column specifications.
Warning message:
“Length of logical index must be 1 or 6064, not 6063”

[1] "No eyetracking data for pl04 (or excluded), skipping..."
[1] "No eyetracking data for pl05 (or excluded), skipping..."
[1] "processing results/pl06_fixations.txt..."

Parsed with column specification:
cols(
  .default = col_character(),
  TRIAL_INDEX = col_double(),
  CURRENT_FIX_ADJUSTED = col_logical(),
  CURRENT_FIX_DURATION = col_double(),
  CURRENT_FIX_END = col_double(),
  CURRENT_FIX_INDEX = col_double(),
  CURRENT_FIX_IS_RT_END = col_logical(),
  CURRENT_FIX_MSG_COUNT = col_double(),
  CURRENT_FIX_PUPIL = col_double(),
  CURRENT_FIX_START = col_double(),
  CURRENT_FIX_X = col_double(),
  CURRENT_FIX_X_RESOLUTION = col_double(),
  CURRENT_FIX_Y = col_double(),
  CURRENT_FIX_Y_RESOLUTION = col_double(),
  IP_END_EVENT_MATCHED = col_logical(),
  IP_END_TIME = col_double(),
  IP_START_EVENT_MATCHED = col_logical(),
  IP_START_TIME = col_double(),
  TRIAL_FIXATION_TOTAL = col_double(),
  TRIAL_START_TIME = col_double(),
  RT = col_double()
  # ... with 8 more columns
)
See spec(...) for full column specifications.
Warning message:
“Length of logical index must be 1 or 13842, not 13841”

length(ed)

Preprocess / Plot Production Data¶

We preprocess the production data at this point because we want to be able to use production scores as a grouping variable in the eyetracking analysis

p_coded = read.csv('productions_coded.csv', stringsAsFactors=F)
print(names(p_coded))
print(nrow(p_coded))

[1] "child"         "itemId"        "Lemma"         "Singular.IPA" 
[5] "Singular.Code" "Plural.IPA"    "Plural.Code"   "Comments"     
[1] 400

codebook = c(
    'No data (0)',
    'No verbal response (1)',
    'Not identifiable/babble (2)',
    'Other word (not noun) (3)',
    'Noun from context (4)',
    'Singular form (5)',
    'Non-conventional plural form (6)',
    'Plural form, approximate articulation (7)',
    'Plural form, adult-like articulation (8)' 
)

codes = data.frame(code = seq(0,8), description = codebook)
codes$description = factor(codes$description, levels = codebook)

# get the item properties into here
pt_ds = read.csv('../PLEARN_experiments/plearn_productiveTask_datasource.csv')
p_coded = merge(p_coded, pt_ds, by.x='Lemma', by.y='s_form')
nrow(p_coded) # the decrease is because the remainder are practice

s_codes = aggregate(itemId ~ Singular.Code + child, p_coded, length)
names(s_codes) = c('code', 'child', 'count')
s_codes = merge(s_codes, codes)
s_codes$s_pl ='singular'
head(s_codes)

p_codes = aggregate(itemId ~ Plural.Code + child, p_coded, length)
names(p_codes) = c('code', 'child','count')
p_codes = merge(p_codes, codes)
p_codes$s_pl = 'plural'
head(p_codes)

code_counts = rbind.fill(s_codes, p_codes)
code_counts$s_pl = factor(code_counts$s_pl, levels =c('singular','plural'))
head(code_counts)

options(repr.plot.width=7, repr.plot.height=10)
ggplot(code_counts) + geom_bar(aes(x=description, y=count, fill=description), stat='identity'
) + theme_bw() +theme(legend.position="none") + theme(axis.text.x = element_text(angle = 45, 
hjust = 1)) + xlab('') + ylab('') + facet_grid(child~s_pl)

# coded trials after  singular productions, #5
p_codes = aggregate(itemId ~ Plural.Code + child, subset(p_coded, Singular.Code == 5), length)
names(p_codes) = c('code', 'child','count')
p_codes = merge(p_codes, codes)
p_codes$type = 'plural'
ggplot(p_codes) + geom_bar(aes(x=description, y=count, fill=description), stat='identity'
) + theme_bw() +theme(legend.position="none") + theme(axis.text.x = element_text(angle = 45, 
hjust = 1)) + xlab('') + ylab(''
) + ggtitle('Responses for Plural Trials after Singular Production') + facet_wrap(~child)

[ ] Sort the two above using the child age

Compute Production Scores¶

# note that both of these are subset to singular succeses
computeProductionScores = function(codes, method = 'narrow'){
    
    if (method == 'narrow'){
        # if they marked the plural in some way,was it adult-like?
        sum(codes %in% c(7,8)) / sum(codes %in% seq(6,8))
    } else if (method == 'broad') {
        # did they mark the plural?
        sum(codes %in% c(7,8)) / sum(codes %in% seq(2,8))
    }
}

narrow_production_scores = aggregate(Plural.Code ~ child, subset(p_coded, Singular.Code == 5), computeProductionScores)
names(narrow_production_scores) = c('participant_name', 'narrow_score')
broad_production_scores = aggregate(Plural.Code ~ child, subset(p_coded, Singular.Code == 5), function(x){computeProductionScores(x,'broad')})
names(broad_production_scores) = c('participant_name','broad_score')

if (!("broad_score" %in% names(subject_info))){
    subject_info = merge(merge(subject_info, narrow_production_scores, by='participant_name', all=T),
    broad_production_scores, all=T, by='participant_name')
}

subset(subject_info, !is.na(narrow_score))

Associate Participant Information With All Records¶

ed_df = do.call('rbind', ed)
#augment ed_df with the properties from subject info
ed_df = merge(ed_df, subject_info, by='filename')
nrow(ed_df)

# add the trial order as a grouping variable
ed_df$first3 = 'remainder'
ed_df$first3[ed_df$TRIAL_INDEX <= 7] = 'first 3 test trials'

Aggregate Eyetracking Timecourses¶

Background / PP Version¶

source('PLEARN_analysis_helper.R')
options(repr.plot.width=6, repr.plot.height=6)
getGroupPlots(ed_df, filter_clause = "type == 'adult' & expt_version == 'scene'",
    mean_pp_duration = mean_pp_duration, group_title = 'Adults in Background Version')

[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + novelty + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + voicing + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + animacystatus + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ label_at_onset"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + label_at_onset"

[1] "facet clause:"
[1] "~ first3"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + first3"

source('PLEARN_analysis_helper.R')
options(repr.plot.width=6, repr.plot.height=6)
getGroupPlots(ed_df, filter_clause = "type == 'child'& expt_version == 'scene'",
    mean_pp_duration = mean_pp_duration, group_title = 'Children in Background Version')

[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target"

Warning message:
“Ignoring unknown parameters: shape”Warning message:
“Removed 1 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + novelty + target"

Warning message:
“Ignoring unknown parameters: shape”Warning message:
“Removed 1 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + voicing + target"

Warning message:
“Ignoring unknown parameters: shape”

Warning message:
“Removed 1 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + animacystatus + target"

Warning message:
“Ignoring unknown parameters: shape”

Warning message:
“Removed 1 rows containing missing values (geom_errorbar).”

source('PLEARN_analysis_helper.R')
options(repr.plot.width=6, repr.plot.height=6)
getGroupPlots(ed_df, filter_clause = "type == 'child' & age_in_days >= 915 & expt_version == 'scene'",
    mean_pp_duration = mean_pp_duration, group_title = 'Older Children (over 30 months) in Background Version')

[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + novelty + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + voicing + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + animacystatus + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ label_at_onset"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + label_at_onset"

[1] "facet clause:"
[1] "~ first3"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + first3"

source('PLEARN_analysis_helper.R')
options(repr.plot.width=6, repr.plot.height=6)
getGroupPlots(ed_df, filter_clause = "type == 'child' & broad_score > .5 & expt_version == 'scene' ",
    mean_pp_duration = mean_pp_duration, group_title = 'Children Producing Plural in Background Version')

[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + novelty + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + voicing + target"

Warning message:
“Ignoring unknown parameters: shape”

Warning message:
“Removed 1 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + animacystatus + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ label_at_onset"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + label_at_onset"

Red/Blue (Simplified) Version¶

source('PLEARN_analysis_helper.R')
options(repr.plot.width=6, repr.plot.height=6)
getGroupPlots(ed_df, filter_clause = "type == 'adult' & expt_version == 'redblue'",
    mean_pp_duration = NULL, group_title = 'Adults in Simplified Version')

[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + novelty + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + voicing + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + animacystatus + target"

Warning message:
“Ignoring unknown parameters: shape”

[1] "facet clause:"
[1] "~ label_at_onset"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + label_at_onset"

[1] "facet clause:"
[1] "~ first3"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + first3"

source('PLEARN_analysis_helper.R')
options(repr.plot.width=6, repr.plot.height=6)
getGroupPlots(ed_df, filter_clause = "type == 'child' & expt_version == 'redblue'",
    mean_pp_duration = NULL, group_title = 'Children in Simplified Version')

[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target"

Warning message:
“Ignoring unknown parameters: shape”Warning message:
“Removed 2 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + novelty + target"

Warning message:
“Ignoring unknown parameters: shape”Warning message:
“Removed 4 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + voicing + target"

Warning message:
“Ignoring unknown parameters: shape”

Warning message:
“Removed 5 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + animacystatus + target"

Warning message:
“Ignoring unknown parameters: shape”

Warning message:
“Removed 4 rows containing missing values (geom_errorbar).”

source('PLEARN_analysis_helper.R')
getGroupPlots(ed_df, filter_clause = "type == 'child' & age_in_days >= 915 & expt_version == 'redblue'",
    mean_pp_duration = NULL, group_title = 'Older Children (over 30 months) in Simplified Version')

[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target"

Warning message:
“Ignoring unknown parameters: shape”Warning message:
“Removed 1 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + novelty + target"

Warning message:
“Ignoring unknown parameters: shape”Warning message:
“Removed 4 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + voicing + target"

Warning message:
“Ignoring unknown parameters: shape”

Warning message:
“Removed 4 rows containing missing values (geom_errorbar).”

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + animacystatus + target"

Warning message:
“Ignoring unknown parameters: shape”

Warning message:
“Removed 3 rows containing missing values (geom_errorbar).”

Comparison Between Experiments¶

getGroupPlot(ed_df, 
        grouping_var = 'expt_version',
        filter_clause = "type == 'adult'",             
        facet_clause = '~ target',
        facet_type = 'wrap',
        loessSpan=.2, 
        x_start = -2000,
        x_end = 3000,
        mean_pp_duration= NULL,
        delay_ms= 367, 
        group_title = "Comparison of Experiment Versions")

[1] "facet clause:"
[1] "~ target"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + expt_version + target"

Warning message:
“Ignoring unknown parameters: shape”

getGroupPlot(ed_df, 
        grouping_var = 'target',
        filter_clause = "type == 'adult'",             
        facet_clause = '~ expt_version',
        facet_type = 'wrap',
        loessSpan=.2, 
        x_start = -2000,
        x_end = 3000,
        mean_pp_duration= NULL,
        delay_ms= 367, 
        group_title = "Comparison of Experiment Versions")

[1] "facet clause:"
[1] "~ expt_version"
[1] "Final aggregate statement is:"
[1] "cfial_bin ~ Time + target + expt_version"

Warning message:
“Ignoring unknown parameters: shape”

Participant Receptive Scores¶

# [ ] Need to update the logic for yoked pair / difference scores

names(ed_df)

source('PLEARN_analysis_helper.R')
by_participant = split(ed_df, ed_df$filename)
normalizeMethods = c('none','yoked','preceding')
all_scores = do.call('rbind', lapply(normalizeMethods, function(normalizeMethod){
    do.call('rbind', lapply(by_participant, function(fixbin_df){
        test_participant_receptive_knowledge(fixbin_df, normalizeMethod, end_analysis_window= 2500)
    }))    
}))

options(repr.plot.width=7, repr.plot.height=5)
ggplot(subset(all_scores, normalizeMethod == 'none')) + geom_violin(
aes(y=prop_looks_to_target, x = participant_type,
colour = participant_type, fill = participant_type), alpha=.25) +
geom_point(aes(x=participant_type, y=prop_looks_to_target), colour ='black',
shape = 4, position=position_jitter(height=.03, width=0)) + facet_wrap(
~partition_name) + theme_bw() + geom_hline(yintercept=.5, 
linetype = 'dashed', colour='red') + ggtitle('Raw Proportion Looks to Target'
) + xlab('Participant Type') + ylab('Proportion Looks to Target')

options(repr.plot.width=7, repr.plot.height=5)
ggplot(subset(all_scores, normalizeMethod == 'yoked')) + geom_violin(
aes(y=prop_looks_to_target, x = participant_type,
colour = participant_type, fill = participant_type), alpha=.25) +
geom_point(aes(x=participant_type, y=prop_looks_to_target), colour ='black',
shape = 4, position=position_jitter(height=.03, width=0)) + facet_wrap(
~partition_name) + theme_bw() + geom_hline(yintercept=0, 
linetype = 'dashed', colour='red') + ggtitle('Yoked Trials'
) + xlab('Participant Type') + ylab('Increase in Target Looking (yoked trials)')

options(repr.plot.width=7, repr.plot.height=5)
ggplot(subset(all_scores, normalizeMethod == 'preceding')) + geom_violin(
aes(y=prop_looks_to_target, x = participant_type,
colour = participant_type, fill = participant_type), alpha=.25) +
geom_point(aes(x=participant_type, y=prop_looks_to_target), position=position_jitter(
height=.03, width=0), colour ='black', shape = 4) + facet_wrap(
~partition_name) + theme_bw() + geom_hline(yintercept=0, 
linetype = 'dashed', colour='red') + ggtitle('Adjusted for Baseline Preference'
) + xlab('Participant Type') + ylab('Increased in Target Looking (vs. preceding 2367 ms)')

Expressive Scores vs. Receptive Scores¶

subject_info[12,]

child_receptive_scores = subset(all_scores, normalizeMethod == 'yoked' & contrast_type == '1-way')
exre = merge(subject_info, child_receptive_scores, by='participant_name')
head(exre)

options(repr.plot.width=6, repr.plot.height=6)
ggplot(exre) + geom_hline(yintercept=0,  color = 'red', linetype = 'dashed') + geom_jitter(
aes(x=broad_score, y=prop_looks_to_target, colour=age_in_months)
) + scale_colour_continuous(type = "viridis") + theme_classic() + ggtitle('Yoked Trials')

Warning message:
“Removed 17 rows containing missing values (geom_point).”

child_receptive_scores = subset(all_scores, normalizeMethod == 'none' & contrast_type == '1-way')
exre = merge(subject_info, child_receptive_scores, by='participant_name')
head(exre)

options(repr.plot.width=6, repr.plot.height=6)
ggplot(exre) + geom_hline(yintercept=.5,  color = 'red', linetype = 'dashed') + geom_jitter(
aes(x=broad_score, y=prop_looks_to_target, colour=age_in_months)
) + scale_colour_continuous(type = "viridis") + theme_classic() + ggtitle('No normalization')

Warning message:
“Removed 17 rows containing missing values (geom_point).”

child_receptive_scores = subset(all_scores, normalizeMethod == 'preceding' & contrast_type == '1-way')
exre = merge(subject_info, child_receptive_scores, by='participant_name')
head(exre)

options(repr.plot.width=6, repr.plot.height=6)
ggplot(exre) + geom_hline(yintercept=0,  color = 'red', linetype = 'dashed') + geom_jitter(
aes(x=broad_score, y=prop_looks_to_target, colour=age_in_months)
) + scale_colour_continuous(type = "viridis") + theme_classic() + ggtitle('Normalized by Preceding')

Warning message:
“Removed 17 rows containing missing values (geom_point).”

Logistic Regression Models for Production¶

library('lme4')

p_coded$s_success = NA 
p_coded$s_success[p_coded$Singular.Code %in% c(5)] = 1
p_coded$s_success[p_coded$Singular.Code %in% seq(1,4)] = 0
p_coded$pl_success = NA 
p_coded$pl_success[p_coded$Plural.Code %in% c(7,8)] = 1
p_coded$pl_success[p_coded$Plural.Code %in% seq(1,6)] = 0

subset(p_coded, pl_success == T)

Singular Production¶

table(p_coded$s_success)

  0   1 
 66 240

library('optimx')

lm = glmer(s_success~novelty+ voicing + animacyStatus + 
    (novelty + voicing + animacyStatus |child) +(1|itemId), family=binomial, data=p_coded,
 control = glmerControl(optimizer="bobyqa",optCtrl=list(maxfun=100000)))
summary(lm)

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: s_success ~ novelty + voicing + animacyStatus + (novelty + voicing +  
    animacyStatus | child) + (1 | itemId)
   Data: p_coded
Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))

     AIC      BIC   logLik deviance df.resid 
   235.8    291.7   -102.9    205.8      291 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-4.1872  0.0490  0.1441  0.2865  2.5895 

Random effects:
 Groups Name                   Variance Std.Dev. Corr             
 child  (Intercept)            5.3423   2.311                     
        noveltynovel           1.3357   1.156    -0.22            
        voicingvoiceless       1.1780   1.085     0.65 -0.11      
        animacyStatusinanimate 0.4212   0.649     0.53  0.34  0.90
 itemId (Intercept)            0.0000   0.000                     
Number of obs: 306, groups:  child, 20; itemId, 16

Fixed effects:
                       Estimate Std. Error z value Pr(>|z|)   
(Intercept)              2.5482     0.8971   2.840   0.0045 **
noveltynovel            -0.9003     0.8107  -1.111   0.2668   
voicingvoiceless         0.9322     0.7729   1.206   0.2278   
animacyStatusinanimate   0.4509     0.6672   0.676   0.4991   
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
            (Intr) nvltyn vcngvc
noveltynovl -0.556              
voicngvclss -0.064 -0.026       
anmcySttsnn -0.196  0.040  0.151

Plural Production¶

table(subset(p_coded, Singular.Code ==5)$pl_success)

  0   1 
142  95

lm = glmer(pl_success~novelty+ voicing + animacyStatus + 
    (novelty+ voicing + animacyStatus|child) +(1|itemId), family=binomial, 
    data=subset(p_coded, Singular.Code ==5),
 control = glmerControl(optimizer="bobyqa",optCtrl=list(maxfun=100000)))

summary(lm)

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: pl_success ~ novelty + voicing + animacyStatus + (novelty + voicing +  
    animacyStatus | child) + (1 | itemId)
   Data: subset(p_coded, Singular.Code == 5)
Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))

     AIC      BIC   logLik deviance df.resid 
   266.0    318.1   -118.0    236.0      222 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.7224 -0.4956 -0.1644  0.4877  2.3454 

Random effects:
 Groups Name                   Variance Std.Dev. Corr          
 child  (Intercept)            1.2191   1.1041                 
        noveltynovel           1.6924   1.3009   0.95          
        voicingvoiceless       0.2530   0.5030   1.00 0.97     
        animacyStatusinanimate 0.7404   0.8605   0.77 0.53 0.73
 itemId (Intercept)            0.0000   0.0000                 
Number of obs: 237, groups:  child, 19; itemId, 16

Fixed effects:
                       Estimate Std. Error z value Pr(>|z|)  
(Intercept)            -0.57688    0.43135  -1.337   0.1811  
noveltynovel           -1.30032    0.61887  -2.101   0.0356 *
voicingvoiceless       -0.05586    0.41457  -0.135   0.8928  
animacyStatusinanimate  0.04004    0.45521   0.088   0.9299  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
            (Intr) nvltyn vcngvc
noveltynovl  0.144              
voicngvclss -0.231  0.111       
anmcySttsnn -0.168  0.049  0.111

Age vs. Plural Knowledge¶

code_counts_with_age = merge(subject_info, code_counts, by.x='id', by.y='child')
code_counts_with_age = merge(code_counts_with_age, codes)
head(code_counts_with_age)

subset(code_counts_with_age, age_in_months > 35 & age_in_months < 38)

options(repr.plot.width=6, repr.plot.height=6)
ggplot(code_counts_with_age) + geom_bar(aes(x=age_in_months, y=count, fill=description),
stat="identity",  width = 0.4, position="fill") + scale_fill_brewer(palette="Set3"
) + facet_wrap(~s_pl, nrow=2) + xlab('Age in Months') + ylab('Count'
) + theme_classic()

Warning message:
“position_stack requires non-overlapping x intervals”Warning message:
“position_stack requires non-overlapping x intervals”

Compare with existing studies:

Zapf and Smith (2007):
Zapf and Ettlinger (2011): for 22-35 months (m=25.8): 52% correct plurals; 33% singulars, remainder are non-answers Experimenter says "Can you tell Teddy to get __" and point to the singulars or the plurals. Common familiar nouns. Teddy bear picks up things or not, i.e. there is reinforcement learning.

Vocabulary vs. Plural Knowledge¶

# pull in CDI scores

Response Time When Looking at Distractor¶

metadata_cols = c('animacystatus', 'background_context','distractorimage',
'distractorloc', 'expt_index', 'full_carrier', 'novel_id', 'novelcolor',
'novelty', 'order', 'pl_carrier', 'pl_form', 'pl_punct', 'pl_sentence',
'pl_wav_index', 'practice', 's_carrier', 's_form', 's_punct',
's_sentence', 's_wav_index', 'simple_pl_stim', 'simple_s_stim', 'target',
'targetimage', 'targetloc', 'targetside', 'test_index', 'trial', 'voicing', 'vowel_carrier',
'RECORDING_SESSION_LABEL')

names(ed_df)

fixreports = split(ed_df, ed_df$filename)

source('PLEARN_analysis_helper.R')
test_df = do.call('rbind',lapply(fixreports, function(fixreport){
    getParticipantRTs(380, fixreport, label_colname="CURRENT_FIX_INTEREST_AREA_LABEL", metadata_cols, include_non_roi_label=T)
}))

Error in if (last_fix[[label_colname]] == ".") {: missing value where TRUE/FALSE needed
Traceback:

1. do.call("rbind", lapply(fixreports, function(fixreport) {
 .     getParticipantRTs(380, fixreport, label_colname = "CURRENT_FIX_INTEREST_AREA_LABEL", 
 .         metadata_cols, include_non_roi_label = T)
 . }))
2. lapply(fixreports, function(fixreport) {
 .     getParticipantRTs(380, fixreport, label_colname = "CURRENT_FIX_INTEREST_AREA_LABEL", 
 .         metadata_cols, include_non_roi_label = T)
 . })
3. FUN(X[[i]], ...)
4. getParticipantRTs(380, fixreport, label_colname = "CURRENT_FIX_INTEREST_AREA_LABEL", 
 .     metadata_cols, include_non_roi_label = T)   # at line 3 of file <text>
5. do.call("rbind", lapply(gaze_trials, function(x) {
 .     getTrialRT(x, delay_ms, label_colname, metadata_cols, buffer_ms, 
 .         include_non_roi_label, target_labels, non_target_labels)
 . }))
6. lapply(gaze_trials, function(x) {
 .     getTrialRT(x, delay_ms, label_colname, metadata_cols, buffer_ms, 
 .         include_non_roi_label, target_labels, non_target_labels)
 . })
7. FUN(X[[i]], ...)
8. getTrialRT(x, delay_ms, label_colname, metadata_cols, buffer_ms, 
 .     include_non_roi_label, target_labels, non_target_labels)

# 10/10/19: this is failing because of an edge case where we don't see any fixations before disambiguation

getRTComparisonPlot = function(test_df, split_var, facet_var = NULL){
    
    agg_equation = paste("rt ~", split_var)
    if (!is.null(facet_var)){
        agg_equation = paste(agg_equation, "+", facet_var)
    }
    summary_df = do.call(data.frame, aggregate(as.formula(agg_equation), data = test_df, FUN = function(x)
        c(mean = mean(x), sd = sd(x))))
    print(names(summary_df))                   
    
    options(repr.plot.width=4, repr.plot.height=4)
    p1 = ggplot(test_df) + geom_violin(mapping=aes(y=rt, x=factor(test_df[[split_var
]]), colour=factor(test_df[[split_var]]))) + theme_bw(
) + geom_errorbar(data = summary_df, mapping=aes(x=factor(summary_df[[split_var
]]), ymin =  rt.mean - rt.sd, ymax = rt.mean + rt.sd), width = 0.1, alpha=.5) + geom_errorbar(data = summary_df, mapping=aes(x=factor(summary_df[[split_var
]]), ymin =  rt.mean, ymax = rt.mean)) + theme_bw(
) + geom_hline(yintercept=0, colour='black', linetype='dashed'
) + geom_hline(yintercept=367, colour='forestgreen') + coord_flip(
ylim= c(0,2000)) + xlab(split_var) + ylab(''
) + theme(legend.position="none")
    
    if (!is.null(facet_var)){
        options(repr.plot.width=4, repr.plot.height=8)
        p1 = p1 + facet_wrap(as.formula(paste('~', facet_var, sep='')), ncol=1)
    }
    print(p1)
                                               
    p2 = ggplot(test_df) + geom_histogram(mapping=aes(x=rt, fill=factor(test_df[[split_var]]))
    ) + theme_bw() + geom_errorbarh(data = summary_df, 
    mapping=aes(y=factor(summary_df[[split_var]]), xmin =  rt.mean - rt.sd, xmax = rt.mean + rt.sd),
    width = 0.1, alpha=.5) + geom_errorbarh(data = summary_df, mapping=aes(y=factor(summary_df[[split_var
    ]]), xmin =  rt.mean, xmax = rt.mean)) + theme_bw() + geom_vline(xintercept=0, 
    colour='black', linetype='dashed') + geom_vline(xintercept=367, colour='forestgreen'
    ) + coord_cartesian(xlim= c(0,2000)) + theme(legend.position="none")
    
    if (!is.null(facet_var)){
        options(repr.plot.width=4, repr.plot.height=8)
        p2 = p2 + facet_wrap(as.formula(paste('~', facet_var, '+', split_var, sep='')), ncol=1)
    } else {
        p2 = p2 + facet_wrap(as.formula(paste('~', split_var, sep='')), ncol=1)
    }
    #print(p2)                                  
                                               
                                               
}

names(test_df)

getRTComparisonPlot(test_df, split_var = 'RECORDING_SESSION_LABEL', facet_var='target')

getRTComparisonPlot(test_df, split_var = 'target')

getRTComparisonPlot(test_df, split_var = 'target', facet_var = 'RECORDING_SESSION_LABEL')

getRTComparisonPlot(test_df,split_var = 'voicing', facet_var='target')

getRTComparisonPlot(test_df,split_var = 'novelty', facet_var='target')

getRTComparisonPlot(test_df,split_var = 'animacystatus', facet_var='target')

names(test_df)

library('lme4')
lm = lmer(rt ~ novelty + voicing + animacystatus + 
    (novelty + voicing + animacystatus | RECORDING_SESSION_LABEL) +(1|s_form), data=
 test_df, )
summary(lm)

Preference Before Disambiguation¶

ed_df$participant_name = ed_df$participant_name.x

beforeafter_disambig_df = subset(ed_df, 
    CURRENT_FIX_INTEREST_AREA_LABEL %in% c('TARGET','DISTRACTOR')) 
beforeafter_disambig_df$looking_at_plural = 0
beforeafter_disambig_df$looking_at_plural[
    beforeafter_disambig_df$CURRENT_FIX_INTEREST_AREA_LABEL == 'TARGET' & 
    beforeafter_disambig_df$target == 'pl'] = 1
beforeafter_disambig_df$looking_at_plural[
    beforeafter_disambig_df$CURRENT_FIX_INTEREST_AREA_LABEL == 'DISTRACTOR' & 
    beforeafter_disambig_df$target == 's'] = 1
beforeafter_disambig_df$beforeafter = 'After Disambiguation'
beforeafter_disambig_df$beforeafter[beforeafter_disambig_df$Time < 367] = 'Before Disambiguation'
beforeafter_disambig_preference = aggregate(looking_at_plural ~ participant_name +
    type + beforeafter + TRIAL_INDEX, 
    beforeafter_disambig_df, mean)

beforeafter_disambig_by_subject = do.call(data.frame, aggregate(looking_at_plural ~ participant_name + type +  beforeafter, beforeafter_disambig_preference, FUN = function(x){c(mean=mean(x), sd = sd(x))}))                                       

print(beforeafter_disambig_by_subject)

            participant_name  type           beforeafter looking_at_plural.mean
1        adultpl12_fixations adult  After Disambiguation              0.5042678
2        adultpl13_fixations adult  After Disambiguation              0.5158088
3        adultpl14_fixations adult  After Disambiguation              0.5038737
4        adultpl15_fixations adult  After Disambiguation              0.4857293
5        adultpl16_fixations adult  After Disambiguation              0.4912726
6             pl06_fixations adult  After Disambiguation              0.5088005
7             pl07_fixations adult  After Disambiguation              0.4915883
8             pl08_fixations adult  After Disambiguation              0.5117331
9  plearn_adults_plad01.xlsx adult  After Disambiguation              0.4974054
10 plearn_adults_plad02.xlsx adult  After Disambiguation              0.4854233
11 plearn_adults_plad03.xlsx adult  After Disambiguation              0.4978965
12 plearn_adults_plad04.xlsx adult  After Disambiguation              0.5064163
13 plearn_adults_plad05.xlsx adult  After Disambiguation              0.4969425
14 plearn_adults_plad06.xlsx adult  After Disambiguation              0.5138653
15 plearn_adults_plad07.xlsx adult  After Disambiguation              0.4843438
16 plearn_adults_plad08.xlsx adult  After Disambiguation              0.4996597
17 plearn_adults_plad09.xlsx adult  After Disambiguation              0.4808543
18 plearn_adults_plad10.xlsx adult  After Disambiguation              0.5243187
19 plearn_adults_plad11.xlsx adult  After Disambiguation              0.4992369
20 plearn_adults_plad12.xlsx adult  After Disambiguation              0.5062847
21 plearn_adults_plad13.xlsx adult  After Disambiguation              0.4773936
22 plearn_adults_plad14.xlsx adult  After Disambiguation              0.4783569
23 plearn_adults_plad15.xlsx adult  After Disambiguation              0.4832023
24         smflip3_fixreport adult  After Disambiguation              0.5012732
25            pl00_fixations child  After Disambiguation              0.5537850
26            pl01_fixations child  After Disambiguation              0.5799120
27            pl03_fixations child  After Disambiguation              0.4929007
28            pl09_fixations child  After Disambiguation              0.5663096
29            pl10_fixations child  After Disambiguation              0.5438719
30            pl11_fixations child  After Disambiguation              0.5520875
31       pl13_fixations.xlsx child  After Disambiguation              0.6219557
32       pl14_fixations.xlsx child  After Disambiguation              0.4986832
33       pl15_fixations.xlsx child  After Disambiguation              0.5590291
34       pl16_fixations.xlsx child  After Disambiguation              0.4227523
35       pl17_fixations.xlsx child  After Disambiguation              0.4051303
36       pl18_fixations.xlsx child  After Disambiguation              0.6106546
37       pl19_fixations.xlsx child  After Disambiguation              0.5237025
38       pl20_fixations.xlsx child  After Disambiguation              0.5341703
39       pl22_fixations.xlsx child  After Disambiguation              0.5854573
40       pl23_fixations.xlsx child  After Disambiguation              0.3910712
41       pl24_fixations.xlsx child  After Disambiguation              0.5359119
42       pl25_fixations.xlsx child  After Disambiguation              0.6123308
43       pl27_fixations.xlsx child  After Disambiguation              0.4784586
44       pl28_fixations.xlsx child  After Disambiguation              0.4945517
45       pl29_fixations.xlsx child  After Disambiguation              0.4958956
46       pl30_fixations.xlsx child  After Disambiguation              0.5752203
47       pl31_fixations.xlsx child  After Disambiguation              0.5076421
48       pl33_fixations.xlsx child  After Disambiguation              0.5272493
49       pl36_fixations.xlsx child  After Disambiguation              0.5188297
50       adultpl12_fixations adult Before Disambiguation              0.4697495
51       adultpl13_fixations adult Before Disambiguation              0.5649460
52       adultpl14_fixations adult Before Disambiguation              0.5002115
53       adultpl15_fixations adult Before Disambiguation              0.5110441
54       adultpl16_fixations adult Before Disambiguation              0.5283740
55            pl06_fixations adult Before Disambiguation              0.5043211
56            pl07_fixations adult Before Disambiguation              0.4643066
57            pl08_fixations adult Before Disambiguation              0.5436851
58 plearn_adults_plad01.xlsx adult Before Disambiguation              0.5723266
59 plearn_adults_plad02.xlsx adult Before Disambiguation              0.5809147
60 plearn_adults_plad03.xlsx adult Before Disambiguation              0.5726449
61 plearn_adults_plad04.xlsx adult Before Disambiguation              0.6057335
62 plearn_adults_plad05.xlsx adult Before Disambiguation              0.5753207
63 plearn_adults_plad06.xlsx adult Before Disambiguation              0.5686803
64 plearn_adults_plad07.xlsx adult Before Disambiguation              0.5029924
65 plearn_adults_plad08.xlsx adult Before Disambiguation              0.5299663
66 plearn_adults_plad09.xlsx adult Before Disambiguation              0.5544098
67 plearn_adults_plad10.xlsx adult Before Disambiguation              0.4995726
68 plearn_adults_plad11.xlsx adult Before Disambiguation              0.5467651
69 plearn_adults_plad12.xlsx adult Before Disambiguation              0.5508098
70 plearn_adults_plad13.xlsx adult Before Disambiguation              0.5646364
71 plearn_adults_plad14.xlsx adult Before Disambiguation              0.5647836
72 plearn_adults_plad15.xlsx adult Before Disambiguation              0.4929646
73         smflip3_fixreport adult Before Disambiguation              0.5480633
74            pl00_fixations child Before Disambiguation              0.5555403
75            pl01_fixations child Before Disambiguation              0.5768089
76            pl03_fixations child Before Disambiguation              0.6547316
77            pl09_fixations child Before Disambiguation              0.6281174
78            pl10_fixations child Before Disambiguation              0.5447676
79            pl11_fixations child Before Disambiguation              0.4789075
80       pl13_fixations.xlsx child Before Disambiguation              0.5392988
81       pl14_fixations.xlsx child Before Disambiguation              0.5291387
82       pl15_fixations.xlsx child Before Disambiguation              0.5330323
83       pl16_fixations.xlsx child Before Disambiguation              0.5603073
84       pl17_fixations.xlsx child Before Disambiguation              0.6219157
85       pl18_fixations.xlsx child Before Disambiguation              0.5180734
86       pl19_fixations.xlsx child Before Disambiguation              0.4557542
87       pl20_fixations.xlsx child Before Disambiguation              0.5566214
88       pl22_fixations.xlsx child Before Disambiguation              0.5243279
89       pl23_fixations.xlsx child Before Disambiguation              0.5581646
90       pl24_fixations.xlsx child Before Disambiguation              0.5651314
91       pl25_fixations.xlsx child Before Disambiguation              0.5152689
92       pl27_fixations.xlsx child Before Disambiguation              0.5556331
93       pl28_fixations.xlsx child Before Disambiguation              0.5393288
94       pl29_fixations.xlsx child Before Disambiguation              0.5397959
95       pl30_fixations.xlsx child Before Disambiguation              0.5626772
96       pl31_fixations.xlsx child Before Disambiguation              0.6105334
97       pl33_fixations.xlsx child Before Disambiguation              0.4971769
98       pl36_fixations.xlsx child Before Disambiguation              0.5048699
   looking_at_plural.sd
1            0.44475922
2            0.35949535
3            0.40782757
4            0.45338155
5            0.43461512
6            0.46628502
7            0.44128874
8            0.44219022
9            0.46179769
10           0.45288550
11           0.41637784
12           0.43606253
13           0.46023293
14           0.45134128
15           0.42252756
16           0.47123758
17           0.43287710
18           0.43336669
19           0.46669849
20           0.45221232
21           0.45181992
22           0.45093683
23           0.43873204
24           0.33552290
25           0.31601546
26           0.17936839
27           0.25915815
28           0.35293079
29           0.29721194
30           0.25665939
31           0.27805795
32           0.23435535
33           0.22795547
34           0.22551579
35           0.30111990
36           0.23237862
37           0.22925241
38           0.16761586
39           0.18150663
40           0.23736101
41           0.28331755
42           0.29045196
43           0.24529988
44           0.17844125
45           0.17384779
46           0.26847287
47           0.21719511
48           0.26540410
49           0.18612437
50           0.18443812
51           0.19310621
52           0.32728089
53           0.24826077
54           0.15221514
55           0.20155161
56           0.17382636
57           0.20889243
58           0.16306815
59           0.19728917
60           0.17364827
61           0.32945754
62           0.09867671
63           0.19981922
64           0.21719277
65           0.36041767
66           0.20928538
67           0.24459995
68           0.33726588
69           0.23243691
70           0.22045906
71           0.11754280
72           0.22049410
73           0.21290987
74           0.30670984
75           0.21348128
76           0.20539461
77           0.29017982
78           0.30462036
79           0.27678600
80           0.21383343
81           0.22753340
82           0.19577386
83           0.17308264
84           0.26329651
85           0.20352239
86           0.22948369
87           0.13067776
88           0.18010164
89           0.28477282
90           0.27484083
91           0.29908800
92           0.26456090
93           0.19830563
94           0.14928096
95           0.28647065
96           0.24001034
97           0.25872863
98           0.15492586

beforeafter_disambig_by_subject$looking_at_plural_low = 
    beforeafter_disambig_by_subject$looking_at_plural.mean - beforeafter_disambig_by_subject$looking_at_plural.sd
beforeafter_disambig_by_subject$looking_at_plural_high = 
    beforeafter_disambig_by_subject$looking_at_plural.mean + beforeafter_disambig_by_subject$looking_at_plural.sd

options(repr.plot.width=4, repr.plot.height=4)
ggplot(subset(beforeafter_disambig_by_subject, beforeafter == "Before Disambiguation" &
type == 'adult')) + geom_errorbar(aes(x=participant_name, ymin= looking_at_plural_low,
 ymax= looking_at_plural_high), color='red') + geom_point(aes(x=participant_name, 
y=looking_at_plural.mean)) + theme_classic() + geom_hline(yintercept=.5, linetype = 'dashed'
) + xlab('Participant') + ylab('Mean Proportion Looks to Plural\n Before Disambiguation (SD)'
) + theme(axis.text.x = element_text(angle = 90, hjust = 1)) + facet_wrap(~type)

options(repr.plot.width=4, repr.plot.height=4)
ggplot(subset(beforeafter_disambig_by_subject, beforeafter == "Before Disambiguation" &
type == 'child')) + geom_errorbar(aes(x=participant_name, ymin= looking_at_plural_low,
 ymax= looking_at_plural_high), color='red') + geom_point(aes(x=participant_name, 
y=looking_at_plural.mean)) + theme_classic() + geom_hline(yintercept=.5, linetype = 'dashed'
) + xlab('Participant') + ylab('Mean Proportion Looks to Plural\n Before Disambiguation (SD)'
) + theme(axis.text.x = element_text(angle = 90, hjust = 1)) + facet_wrap(~type)

# should the last fix before disambig be excluded
#[X] Confirm that 0 is the point of disambig for everyone
#[X] propagate the information about the trials
#[X] Make sure that my ordering is doing the right thing
#[X] how to account for the 200 ms exclusion
#[X] have a bit of a buffer on either side of 0 in case there’s an NA instead of a 0/1 in that exact timebin
#    -- so far this is just looking backwards
#[ ] need to recompute wrt the beginning of the /s/ or /z/ segment, not the end

Experimental Visualizations + Analyses¶

Binomial Test for Receptive Knowledge¶

# requires updating since refactoring

pl06_fixbins$is_looking_at_target  = as.numeric(pl06_fixbins$CURRENT_FIX_INTEREST_AREA_LABEL == 'TARGET')

rlist = getPlotForMethod(fixbin_dfs, adult_fixbin_dfs, 'none')
options(repr.plot.width=10, repr.plot.height=5)
print(rlist$p1)
ggsave('figures/binomialTest_nopreprocessing.pdf', width=20, height=10)

print(rlist$p2)

False negative rate around .1 on 2-way contrasts when halfway between adult-like knowledge and zero knowledge

False positve rate of around .02 on 1-way contrasts when zero-knowledge

# takes a long time to run because of the bootstrapping
sim_yoked = getPlotForMethod(fixbin_dfs,adult_fixbin_dfs, 'yoked')
options(repr.plot.width=10, repr.plot.height=5)
print(sim_yoked$p1)
ggsave('figures/binomialTest_yoked.pdf', width=20, height=10)

sim_yoked$p2

# takes a long time to run because of the bootstrapping
sim_preceding = getPlotForMethod(fixbin_dfs,adult_fixbin_dfs, 'preceding')
options(repr.plot.width=10, repr.plot.height=5)
print(sim_preceding$p1)
ggsave('figures/binomialTest_preceding.pdf', width=20, height=10)

sim_preceding$p2

Gaze Stickiness Plot¶

fixreport_path = 'results/smflip3_fixreport.txt'
participant_name = gsub('.txt','',tail(strsplit(fixreport_path, '/')[[1]]))[2]

gaze = blabr::fixations_report(fixreport_path)
gaze = merge(gaze, audio_timings[,c('audiotarget','disambig_time')])
gaze$CURRENT_FIX_END = gaze$CURRENT_FIX_END - gaze$disambig_time
gaze$CURRENT_FIX_START = gaze$CURRENT_FIX_START - gaze$disambig_time

fixbins = binifyFixations(gaze, keepCols=c("RECORDING_SESSION_LABEL",
"CURRENT_FIX_INTEREST_AREA_LABEL",
"RT",
"expt_index",
"target",
"s_form",
"novelty",
"animacystatus",
"voicing"))
subset(fixbins, Time==3200)

ggplot(subset(fixbins, s_form =='mip')) + geom_point(aes(x=Time, y=CURRENT_FIX_INTEREST_AREA_LABEL)
) + facet_wrap(~s_form * target) + geom_vline(xintercept=0,
	colour='black')

ggplot(subset(sm_fixbins, CURRENT_FIX_INTEREST_AREA_LABEL!='.')) + geom_point(aes(x=Time,
                                                                               y=s_form, colour=s_form)) + facet_wrap(CURRENT_FIX_INTEREST_AREA_LABEL ~ target) + geom_vline(xintercept=0, colour='black')

ggplot(subset(pl00_fixbins, CURRENT_FIX_INTEREST_AREA_LABEL!='.')) + geom_point(aes(x=Time,
    y=s_form, colour=s_form)) + facet_wrap(CURRENT_FIX_INTEREST_AREA_LABEL ~ target) + geom_vline(xintercept=0, colour='black')

ggplot(subset(pl01_fixbins, CURRENT_FIX_INTEREST_AREA_LABEL!='.')) + geom_point(aes(x=Time,
    y=s_form, colour=s_form)) + facet_wrap(CURRENT_FIX_INTEREST_AREA_LABEL ~ target) + geom_vline(xintercept=0, colour='black')

ggplot(subset(pl03_fixbins, CURRENT_FIX_INTEREST_AREA_LABEL!='.')) + geom_point(aes(x=Time,
    y=s_form, colour=s_form)) + facet_wrap(CURRENT_FIX_INTEREST_AREA_LABEL ~ target) + geom_vline(xintercept=0, colour='black')

	Lemma	child	itemId	Singular.IPA	Singular.Code	Plural.IPA	Plural.Code	Comments	novel_id	background_context	⋯	simple_s_stim	animacyStatus	test_index	trial_index	Trial	Practice	NovelColor	book	s_success	pl_success
1	ant	pl20	10	ant	5	two ant	8		!	tree	⋯	Look at the ant!	animate	9	5	12	n	!	2	1	1
2	ant	pl14	10	ant	5	ant(th)	7		!	tree	⋯	Look at the ant!	animate	9	5	12	n	!	2	1	1
5	ant	pl16	10	I don't know what they ant	5	two ants on… two ants… two ant	8		!	tree	⋯	Look at the ant!	animate	9	5	12	n	!	2	1	1
7	ant	pl22	10	ant	5	ants	8	Had a potty accident; take intermission	!	tree	⋯	Look at the ant!	animate	9	5	12	n	!	2	1	1
8	boat	pl20	6	0	0	two boats	8		!	hill	⋯	Look at the boat!	inanimate	13	7	13	n	!	2	NA	1
11	boat	pl10	6	boat splish splash	5	two boats	8		!	hill	⋯	Look at the boat!	inanimate	13	7	13	n	!	2	1	1
12	boat	pl24	6	it supposed to be in the water, boat	5	two boats	8		!	hill	⋯	Look at the boat!	inanimate	13	7	13	n	!	2	1	1
14	boat	pl22	6	boat	5	boats	8		!	hill	⋯	Look at the boat!	inanimate	13	7	13	n	!	2	1	1
16	bug	pl14	1	no….bug	5	bugs… more bugs	8		!	house	⋯	Look at the bug!	animate	1	1	3	n	!	2	1	1
21	bug	pl22	1	buug	5	bugs	8		!	house	⋯	Look at the bug!	animate	1	1	3	n	!	2	1	1
22	cat	pl13	16	cat	5	two cats	8		!	house	⋯	Look at the cat!	animate	10	5	16	n	!	1	1	1
25	cat	pl15	16	cat	5	cats…cat	7		!	house	⋯	Look at the cat!	animate	10	5	16	n	!	1	1	1
31	cat	pl27	16	cat	5	cat ssss	8		!	house	⋯	Look at the cat!	animate	10	5	16	n	!	1	1	1
32	cat	pl17	16	cat	5	cats	8		!	house	⋯	Look at the cat!	animate	10	5	16	n	!	1	1	1
40	cow	pl17	14	cow	5	cows	8		!	house	⋯	Can you find the cow?	animate	2	1	10	n	!	1	1	1
42	cow	pl19	14	cow	5	cows	8	short whistle s	!	house	⋯	Can you find the cow?	animate	2	1	10	n	!	1	1	1
45	cow	pl15	14	caow	5	cow(sh)	7		!	house	⋯	Can you find the cow?	animate	2	1	10	n	!	1	1	1
53	cup	pl27	11	sippy cup	5	sippy cups	8		!	couch	⋯	Can you find the cup?	inanimate	14	7	4	n	!	1	1	1
57	cup	pl13	11	cu	5	two cups	8		!	couch	⋯	Can you find the cup?	inanimate	14	7	4	n	!	1	1	1
58	cup	pl19	11	cup	5	cupssss	8		!	couch	⋯	Can you find the cup?	inanimate	14	7	4	n	!	1	1	1
59	cup	pl23	11	cup	5	cups	8		!	couch	⋯	Can you find the cup?	inanimate	14	7	4	n	!	1	1	1
63	deg	pl16	3	it's too (agrees to hard)	3	two degs	8		2048-600	tree	⋯	Can you find the deg?	animate	17	9	1	n	red	2	0	1
73	duck	pl22	13	duck	5	ducks	8		!	house	⋯	Can you find the duck?	animate	11	6	15	n	!	2	1	1
76	dup	pl14	12	dup	5	dup(th)	7		2028-600	hill	⋯	Look at the dup!	animate	25	13	8	n	blue	2	1	1
78	dup	pl22	12	dup	5	dups	8		2028-600	hill	⋯	Look at the dup!	animate	25	13	8	n	blue	2	1	1
79	dup	pl16	12	I don't know what the dup	5	two dups	8		2028-600	hill	⋯	Look at the dup!	animate	25	13	8	n	blue	2	1	1
82	egg	pl16	16	e(gh)	5	two eggs	8	when asked what the end page says (the end) participant reports "I love you"	!	couch	⋯	Can you find the egg?	inanimate	5	3	10	n	!	2	1	1
85	egg	pl20	16	a egg	5	two eggs	8		!	couch	⋯	Can you find the egg?	inanimate	5	3	10	n	!	2	1	1
87	egg	pl22	16	egg	5	eggs	8	seems like a devoiced s	!	couch	⋯	Can you find the egg?	inanimate	5	3	10	n	!	2	1	1
89	fip	pl27	6	fip	5	fips	8	laughs afterward	2030-600	couch	⋯	Look at the fip!	animate	26	13	12	n	blue	1	1	1
⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋱	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮
233	pog	pl17	1	pog	5	pogs	8		2042-600	tree	⋯	Look at the pog!	inanimate	22	11	13	n	red	1	1	1
240	pog	pl27	1	pog	5	pogs	8		2042-600	tree	⋯	Look at the pog!	inanimate	22	11	13	n	red	1	1	1
243	rabbit	pl19	2	rabbit	5	a rabbits	8	mom encourages louder	!	tree	⋯	Can you find the rabbit?	animate	12	6	14	n	!	1	1	1
250	rabbit	pl13	2	rabbit	5	two rabbits	8		!	tree	⋯	Can you find the rabbit?	animate	12	6	14	n	!	1	1	1
252	rabbit	pl17	2	ribbit	5	rebbets	8		!	tree	⋯	Can you find the rabbit?	animate	12	6	14	n	!	1	1	1
253	rabbit	pl04	2	rabbit	5	two rabbitsch	7		!	tree	⋯	Can you find the rabbit?	animate	12	6	14	n	!	1	1	1
254	rabbit	pl27	2	rabbit	5	rabbits	8		!	tree	⋯	Can you find the rabbit?	animate	12	6	14	n	!	1	1	1
255	spoon	pl20	14	soon	5	soons	8		!	hill	⋯	Can you find the spoon?	inanimate	7	4	7	n	!	2	1	1
256	spoon	pl14	14	soons	8	soons	8		!	hill	⋯	Can you find the spoon?	inanimate	7	4	7	n	!	2	NA	1
260	spoon	pl16	14	a poon. Poon. No no I can turn the pictures	5	poons… two poons	8		!	hill	⋯	Can you find the spoon?	inanimate	7	4	7	n	!	2	1	1
261	spoon	pl22	14	spoon	5	spoons	8		!	hill	⋯	Can you find the spoon?	inanimate	7	4	7	n	!	2	1	1
263	star	pl13	4	star	5	two stars	8		!	couch	⋯	Look at the star!	inanimate	8	4	7	n	!	1	1	1
264	star	pl01	4	ya, star, twinkle twinkle	5	tarsh, tarsh	7		!	couch	⋯	Look at the star!	inanimate	8	4	7	n	!	1	1	1
265	star	pl17	4	sar	5	sars	8	question intonation on the plural	!	couch	⋯	Look at the star!	inanimate	8	4	7	n	!	1	1	1
267	star	pl27	4	staaar	5	staaas	8		!	couch	⋯	Look at the star!	inanimate	8	4	7	n	!	1	1	1
270	star	pl19	4	tar	5	tars	8		!	couch	⋯	Look at the star!	inanimate	8	4	7	n	!	1	1	1
273	star	pl03	4	tar	5	tarsh	7		!	couch	⋯	Look at the star!	inanimate	8	4	7	n	!	1	1	1
275	teb	pl20	8	teb…. Blue, rd, yellow, green	5	two tebs	8		2061-600	couch	⋯	Look at the teb!	inanimate	23	12	14	n	red	2	1	1
278	teb	pl14	8	te ba	5	tebs	8		2061-600	couch	⋯	Look at the teb!	inanimate	23	12	14	n	red	2	1	1
280	teb	pl22	8	teb	5	tebs	8		2061-600	couch	⋯	Look at the teb!	inanimate	23	12	14	n	red	2	1	1
281	teb	pl16	8	no...I don't know… on one	NA	two tep(sh)	7		2061-600	couch	⋯	Look at the teb!	inanimate	23	12	14	n	red	2	NA	1
285	tep	pl27	5	tep	5	teps	8		2005-600	house	⋯	Look at the tep!	inanimate	32	16	8	n	red	1	1	1
286	tep	pl17	5	tep	5	teps	8		2005-600	house	⋯	Look at the tep!	inanimate	32	16	8	n	red	1	1	1
294	tep	pl13	5	tep	5	two teps	8		2005-600	house	⋯	Look at the tep!	inanimate	32	16	8	n	red	1	1	1
297	tig	pl23	13	tig	5	tig tigs	8		2057-600	hill	⋯	Look at the tig!	inanimate	24	12	15	n	blue	1	1	1
301	tig	pl13	13	tig	5	two tigs	8		2057-600	hill	⋯	Look at the tig!	inanimate	24	12	15	n	blue	1	1	1
312	truck	pl27	8	truck	5	truckssssss	8		!	couch	⋯	Look at the truck!	inanimate	16	8	11	n	!	1	1	1
314	truck	pl13	8	truck	5	two trucks	8		!	couch	⋯	Look at the truck!	inanimate	16	8	11	n	!	1	1	1
316	truck	pl19	8	a dump truck	5	dump trucks	8		!	couch	⋯	Look at the truck!	inanimate	16	8	11	n	!	1	1	1
319	truck	pl17	8	truck	5	trucks	8		!	couch	⋯	Look at the truck!	inanimate	16	8	11	n	!	1	1	1

id	include	filename	age_in_days	gender	type	expt_version	comment	participant_name	age_in_months
plsm	1	smflip3_fixreport.txt	NA	M	adult	scene	eyetracking only	smflip3	NA
pl00	1	pl00_fixations.txt	769	F	child	scene		pl00	25.21311
pl01	1	pl01_fixations.txt	788	F	child	scene		pl01	25.83607
pl03	1	pl03_fixations.txt	810	F	child	scene		pl03	26.55738
pl04	1		858	F	child	scene	storybook only; need VNA id to get dob	NA	28.13115
pl05	1		862	F	child	scene	storybook only; need VNA id to get dob	NA	28.26230
pl06	1	pl06_fixations.txt	NA		adult	scene	eyetracking only	pl06	NA
pl07	1	pl07_fixations.txt	NA		adult	scene	eyetracking only	pl07	NA
pl08	1	pl08_fixations.txt	NA		adult	scene	eyetracking only	pl08	NA
pl09	1	pl09_fixations.txt	796	F	child	scene		pl09	26.09836
pl10	1	pl10_fixations.txt	897	M	child	scene		pl10	29.40984
adultpl12	1	adultpl12_fixations.txt	NA		adult	scene		adultpl12	NA
adultpl13	1	adultpl13_fixations.txt	NA		adult	scene		adultpl13	NA
adultpl14	1	adultpl14_fixations.txt	NA		adult	scene		adultpl14	NA
adultpl15	1	adultpl15_fixations.txt	NA		adult	scene		adultpl15	NA
adultpl16	1	adultpl16_fixations.txt	NA		adult	scene		adultpl16	NA
pl11	1	pl11_fixations.txt	792	M	child	scene		pl11	25.96721
pl13	1	pl13_fixations.xlsx	1450	M	child	scene		pl13	47.54098
pl14	1	pl14_fixations.xlsx	1057	M	child	scene		pl14	34.65574
pl15	1	pl15_fixations.xlsx	876	F	child	scene		pl15	28.72131
pl16	1	pl16_fixations.xlsx	1049	M	child	scene		pl16	34.39344
pl17	1	pl17_fixations.xlsx	926	F	child	scene		pl17	30.36066
pl18	1	pl18_fixations.xlsx	1034	F	child	scene		pl18	33.90164
pl19	1	pl19_fixations.xlsx	947	M	child	scene		pl19	31.04918
pl20	1	pl20_fixations.xlsx	1008	F	child	scene		pl20	33.04918
pl21	0	pl21_fixations.xlsx	1100	F	child	redblue	out of order reschedule	pl21	36.06557
pl22	1	pl22_fixations.xlsx	1408	M	child	scene		pl22	46.16393
pl23	1	pl23_fixations.xlsx	850	M	child	scene		pl23	27.86885
pl24	1	pl24_fixations.xlsx	1091	M	child	scene		pl24	35.77049
pl25	1	pl25_fixations.xlsx	1091	M	child	scene		pl25	35.77049
pl26	0	pl26_fixations.xlsx	NA		child	scene	rescheduling	pl26	NA
pl27	1	pl27_fixations.xlsx	1107	F	child	scene		pl27	36.29508
pl28	1	pl28_fixations.xlsx	951	F	child	scene		pl28	31.18033
pl29	1	pl29_fixations.xlsx	931	F	child	scene		pl29	30.52459
pl30	1	pl30_fixations.xlsx	931		child	redblue	first redblue	pl30	30.52459
pl31	1	pl31_fixations.xlsx	740		child	redblue		pl31	24.26230
pl33	1	pl33_fixations.xlsx	725		child	redblue		pl33	23.77049
pl36	1	pl36_fixations.xlsx	1079		child	redblue		pl36	35.37705
plad01	1	plearn_adults_plad01.xlsx	NA		adult	redblue		plearn	NA
plad02	1	plearn_adults_plad02.xlsx	NA		adult	redblue		plearn	NA
plad03	1	plearn_adults_plad03.xlsx	NA		adult	redblue		plearn	NA
plad04	1	plearn_adults_plad04.xlsx	NA		adult	redblue		plearn	NA
plad05	1	plearn_adults_plad05.xlsx	NA		adult	redblue		plearn	NA
plad06	1	plearn_adults_plad06.xlsx	NA		adult	redblue		plearn	NA
plad07	1	plearn_adults_plad07.xlsx	NA		adult	redblue		plearn	NA
plad08	1	plearn_adults_plad08.xlsx	NA		adult	redblue		plearn	NA
plad09	1	plearn_adults_plad09.xlsx	NA		adult	redblue		plearn	NA
plad10	1	plearn_adults_plad10.xlsx	NA		adult	redblue		plearn	NA
plad11	1	plearn_adults_plad11.xlsx	NA		adult	redblue		plearn	NA
plad12	1	plearn_adults_plad12.xlsx	NA		adult	redblue		plearn	NA
plad13	1	plearn_adults_plad13.xlsx	NA		adult	redblue		plearn	NA
plad14	1	plearn_adults_plad14.xlsx	NA		adult	redblue		plearn	NA
plad15	1	plearn_adults_plad15.xlsx	NA		adult	redblue		plearn	NA

code	child	count	description	s_pl
0	pl01	1	No data (0)	singular
0	pl18	4	No data (0)	singular
0	pl24	1	No data (0)	singular
0	pl20	5	No data (0)	singular
1	pl05	14	No verbal response (1)	singular
1	pl15	1	No verbal response (1)	singular

code	child	count	description	s_pl
0	pl01	2	No data (0)	plural
0	pl18	3	No data (0)	plural
0	pl19	1	No data (0)	plural
0	pl20	4	No data (0)	plural
1	pl01	2	No verbal response (1)	plural
1	pl03	2	No verbal response (1)	plural